This proposal will create a center for Large Scale Production of Perturbagen-lnduced Cellular Signatures at Harvard Medical School and collaborating institutions, with a focus on perturbations provoked by small molecule drugs and cellular signatures measured using diverse biochemical and single-cell assays. The result will be a large, self-consistent and diverse set of network-centric Pharmacological Response Signatures that provide unique insight into disease processes, drug mechanism/selectivity and ultimately patient-specific responses to therapy. The initial focus of the Center will be small molecule kinase inhibitors, versatile perturbagens with high translational potential. We will use known inhibitors and also expand dramatically the publicly documented collection of inhibitors through new medicinal chemistry and use of kinome-wide selectivity assays. The responses of a large collection of human tumor cells and some primary cells to kinase inhibitors, will be assayed using multiplex biochemical assays (for 20-100 proteins) involving bead-based sandwich immunoassays and reverse-phase lysate microarrays, and single-cell assays (using imaging and flow cytometry) for cell cycle state, commitment to senescence or apoptosis, mesenchymal vs. epithelial phenotype and markers of primitive (stem-cell) status. Data will be collected, integrated and distributed using a series of novel, interoperable software tools that manipulate semantically-typed data arrays based on a new XML/HDF5 format. A multi-faceted informatics program will link these phenotypic and biochemical measures of cellular response to a rich and growing set of genomic data being collected by others. These goals will be met through pursuit of six linked specific aims. Aim 1 will focus on existing - largely clinical grade - kinase inhibitors and a set of 45 cell lines that are known to display diverse drug responses land for which extensive genomic data are available. Aim 2 will enlarge the set of perturbagens by developing a large library of kinase inhibitors using new and existing chemistry and profiling biochemical specificity across the kinome. Aim 3 will combine existing and novel compounds in a dose-response analysis across a set of >1000 tumor cell lines to identify representative cell lines and outliers which, in Aim 4, will subjected to detailed analysis at a single-cell level. Aims 5-6 will develop and deploy the information processing systems needed to collect, systematize and distribute diverse data types. This will involve a novel set of interoperable software tools that incorporate emerging no-SQL and semantic web concepts. Methods for adaptive experimental design will be developed to focus data collection on those areas of the doseresponse landscape where signatures are most informative. The final product will be a large publiclyavailable data set radically different from, but highly complementary to, the expression profiles and genome data that are the primary focus of current high-throughput biological studies on perturbagen-lnduced cellular signatures.